A tale of two tissues: Probing gene expression in a complex insect‐induced gall

Plant galls are novel and sometimes dramatic plant organs whose development is initiated and controlled by parasitic microbes, nematodes, insects and mites. For arthropods, galls provide relative safety from enemies and abiotic stresses while providing nutrition. Galls are formed entirely by the plant, whose transcriptional pathways are modified and coopted to produce a structure specific to the galler species; they comprise a classic example of Dawkins’ “extended phenotype”. Arthropod‐elicited galls are unique in that they are often anatomically complex (Figure 1a), with multiple differentiated tissue types (Figure 1b). A growing number of investigators have studied changes in hostplant gene expression to understand arthropod gall development. In this issue of Molecular Ecology, Martinson et al. (2021) report using RNA sequencing to explore tissue‐specific gene expression associated with anatomical and functional gall complexity, demonstrating for the first time that gall tissues are as different transcriptionally as they are anatomically.

The gall interior is clearly divided into several major tissue types.
The wasp larva develops in an inner capsule lined with cells on which it feeds, and which is suspended in the hollow outer gall cortex by connecting filaments. Martinson et al. (2021) demonstrate that this gall is complex transcriptionally as well as anatomically. They assembled a de novo transcriptome for Q. rubra from RNA extracted from inner and outer gall and adjacent leaf tissues, then identified genes whose expression differed significantly among them. They found large gene expression changes in the gall, with 28% of oak genes differentially expressed in the gall versus leaf. What is unique in this study is the discovery of similar gene expression differences within a gall.
If the gall is different from a leaf, then what is it? A principal components analysis incorporating additional transcriptional data from other studies found that the outer gall transcriptome resembled that from twigs, leaf buds and reproductive structures but not leaves. In contrast, the inner gall transcriptome was unlike Plant galls are novel and sometimes dramatic plant organs whose development is initiated and controlled by parasitic microbes, nematodes, insects and mites. For arthropods, galls provide relative safety from enemies and abiotic stresses while providing nutrition. Galls are formed entirely by the plant, whose transcriptional pathways are modified and coopted to produce a structure specific to the galler species; they comprise a classic example of Dawkins' "extended phenotype". Arthropod-elicited galls are unique in that they are often anatomically complex ( Figure 1a  Although outer gall tissues are green, Martinson et al. (2021) found expression of photosynthesis-related genes to be reduced in outer gall tissues relative to ungalled leaves, and suppressed or blocked in inner tissues, suggesting that inner gall tissues are effectively heterotrophic. This is compatible with a widespread view that galls are sinks that draw plant metabolites and nutrients from surrounding nongall tissues (Giron et al., 2016), and the few galling species known to negatively impact hostplant fitness may do so by outcompeting naturally occurring sinks (e.g., galling aphids on poplar and Hessian Fly on wheat, Compson et al., 2011;Subramanyam et al., 2021). Elevated expression of genes related to sugar and amino acid metabolism in both outer and inner gall tissues suggests that the entire D. quercuspalustris gall supports transport of plant metabolites to the inner gall tissues via the connecting strands. Biochemical and histological studies have established that cells in the inner gall tissues closest to the feeding insect become specialized "nutritive" cells on which the insect feeds (Ferreira et al., 2019). Martinson et al. (2021) are the first to sample those tissues specifically for gene expression. They demonstrated elevated expression of genes encoding sucrose synthase in nutritive tissues, a suitably localised candidate mechanism for the strong sink effect of the inner gall. Expression of genes characteristic of gluconeogenesis was also greatly enhanced in the inner tissues. Martinson et al. (2021) suggest that glycolysis in these inner tissues might convert pyruvate, lactate or glycerol into glucose to feed the wasp larva. Starch often accumulates near the feeding insect in cynipid galls, but not in the cells actually consumed by the parasite. Interestingly, Martinson et al. (2021) found no differential expression of genes encoding starch degrading enzymes, suggesting that the wasp larva might digest starch on its own. Since galling arthropods attack by chewing or sucking and siphoning resources away from their host, one might expect them to evoke defence responses. To the contrary, Martinson et al. (2021) found expression of defence-related genes to be strikingly suppressed in the inner gall tissue and reduced in the outer gall. The few exceptions appear to be genes encoding proteins targeting microbes, especially fungi, which the authors point out are a major threat to galling insects. One possibility is thus that induced expression of these proteins is a defensive component of the gallwasp's extended phenotype. An alternative is that these proteins are defensive responses against eliciting or suppressing pathogens or gall-inducing symbionts introduced by the galler. While symbionts are essential partners in some insect galls (e.g., fungal symbionts in galls induced by some cecidomyiid midges), neither Martinson et al.  (Schultz et al., 2019), which has its own shortcomings. Martinson et al. (2021) sampled galls at a single time point in their development, so they cannot offer major insights about initiation of gall development or signals that control it. Phytohormones, some of which are produced by insects, are probably gall initiation and developmental signals (Tooker & Helms, 2014). While auxin signalling is central to plant organogenesis and has been implicated in development of other galls (Tooker & Helms, 2014), Martinson et al. (2021 found little evidence of it in this gall. This is probably because the influence of auxin often involves a few cells over a short time span (Galvan-Ampudia et al., 2020). Like many others, Martinson et al. (2021) found upregulated expression of genes in ethylene synthesis and signalling pathways. However, ethylene's functions are many, including defence, growth, and development.
The study by Martinson et al. (2021) is the second to report elevated expression of an ENOD gene in a cynipid gall (Hearn et al., 2019), which together with ethylene guides formation of gall-like rhizobium nodules (Larrainzar et al., 2015). Martinson et al. (2021) also found elevated expression of genes involved in RNA modification, which may control expression of many developmental genes in the gall.
Molecular study of arthropod gall development is the only way we will understand how these remarkable structures come to be. To date, much of the functional inference from transcriptomic data has been based on gene ontology (GO) enrichment. However, this approach is not very informative without determining which genes are up-versus downregulated and what their impact may be in each case (Schultz et al., 2019). The study by Martinson et al. (2021) is one of few to go beyond simply listing enriched GO categories. Functional dissection of the roles of specific genes in gall induction and development is limited by a lack of galls formed on "model" plants (except for Vitis and Populus) with good functional annotation and the ability to modify the expression of key developmental genes over a time course (Betancourt et al., 2020). Relevant molecular and metabolomic resources are nevertheless growing rapidly, and genome editing technologies are increasingly applicable to nonmodel systems (Dickinson et al., 2020).
The study by Martinson et al. (2021) supports many observations made via histology, biochemistry, and a few transcriptome studies: the suppression of photosynthesis and development of a metabolic sink resulting in heterotrophy, suppressed defence responses close to the parasite, the metabolic characteristics of the tissues on which the parasite feeds, and the massive changes in gene expression compared with control tissues. They have, however, for the first time illustrated the spatial array of many of these changes on a transcriptional level. The picture that emerges is of a unique plant organ whose functional and anatomical complexity is matched and probably created by transcriptional changes controlled by the arthropod parasite. Exactly how the parasite achieves this remains unknown.

ACK N OWLED G EM ENTS
We acknowledge support to GNS from UK NERC grant NE/ T000120/1.